aufklarer/WeSpeaker-ResNet34-LM-CoreML

🤗 Hugging Face sourceaudio-classificationmit13 MBother✓ 3 checksumsupdated today

Submit in one command

Run it next to your model folder. It makes the torrent, checks your files against Hugging Face, and submits it. You just start seeding and paste your key from your account. It only reads your files and never changes them. Read the script first if you like.

curl -fsSL https://pirateface.co/package.sh | bash -s -- --repo aufklarer/WeSpeaker-ResNet34-LM-CoreML ./model-folder

Needs a seeder →

WeSpeaker-ResNet34-LM — CoreML

CoreML conversion of WeSpeaker ResNet34-LM for Apple Neural Engine.

Produces 256-dimensional L2-normalized speaker embeddings from audio.

Model Details

Detail	Value
Architecture	ResNet34 with statistics pooling
Parameters	~6.6M
Input	80-bin log-mel spectrogram (16kHz)
Output	256-dim L2-normalized speaker embedding
BatchNorm	Fused into Conv2d at conversion time

Usage

let model = try await WeSpeakerModel.fromPretrained(backend: .coreML)
let embedding = model.embed(audio: samples, sampleRate: 16000)
let similarity = WeSpeakerModel.cosineSimilarity(embeddingA, embeddingB)

Variants